We are asking artificial intelligence to do things it cannot do, and neglecting the one thing it does remarkably well.
It cannot decide for you. It has no responsibility, no consequences to carry, no history with your teams. What it can do is stop you thinking too fast.
The real risk is not delegating the decision
Few executives genuinely hand their trade-offs to a model. The risk is quieter: it is validation. You state your position, you submit it, you receive a well-constructed answer pointing the same way, and you leave reassured.
A language model is built to be helpful and cooperative. Ask it to strengthen your reasoning and it will strengthen it. You have just built yourself a very articulate echo chamber.
The interesting use is the exact opposite. Ask it to take your reasoning apart. To build the strongest case against your decision. To adopt the point of view of whoever has the most to lose.
Three uses that change the quality of thinking
Widening the field of scenarios. An executive spontaneously explores two or three, usually the familiar ones. A machine generates fifteen in three minutes. Twelve will be worthless. One or two will open a door you had not seen.
Structuring a confused situation. When a subject is running in every direction, having it sorted, categorised and ranked often reveals that the problem was not the one you thought.
Preparing a confrontation. Before a negotiation or a difficult committee, have it play your toughest opponent. You find your weak points before he does.
What remains irreducibly human
None of these uses touches what actually makes an executive decision difficult.
What is difficult is not identifying the right option. It is owning it in front of people who will disagree. It is arbitrating between two legitimate loyalties. It is going back on a position you defended publicly.
No model will carry that for you. It has nothing to lose.
Technology and judgement
I trained in artificial intelligence on the Oxford Artificial Intelligence Programme at Saïd Business School, and I use it daily in my work as a sparring partner.
The division of labour seems clear to me: technology to widen the field of the possible, sparring partnership to exercise judgement. One produces options, the other helps you choose.
Confusing the two means risking deciding faster, and worse.
Read also - What you no longer see, and what no one dares tell you · What your non-decisions cost you
Or see the three ways of working together.
Thirty minutes, confidential and without commitment. You set out the situation; I tell you honestly whether I am the right person. Book a 30-minute introductory call →